Agentic AI training
Learn how AI agents work – and how to design them with control
Understand the architecture behind agentic AI: models, tools, retrieval, state, workflows, approvals, evaluation and observability. Build agents around bounded business tasks rather than treating them as autonomous employees.
Who it is for
Agentic AI training for
- AI and automation practitioners
- Data professionals
- Developers and technical analysts
- Solution architects
- Technical managers
- Teams evaluating agentic workflows
Courses
Agentic AI courses
Live online, small groups, hands-on labs.
Agentic AI: Designing Controlled Agents
How AI agents work and how to build them around bounded tasks: tools, retrieval, approvals, evaluation and production patterns.
$229 per learner
What the training covers
Core modules
- Agents vs chatbots vs workflows
- Tool calling
- Prompt and instruction design
- Retrieval and context
- State and memory concepts
- Multi-step orchestration
- Human-in-the-loop approvals
- Permissions and action boundaries
- Evaluation, observability and production patterns
From chatbot to agent
Each step adds capability – and responsibility
- Chatbot: answers questions AI
- Assistant: uses approved knowledge System
- Workflow: follows defined steps Rules
- Agent: chooses tools, within limits AI
- Approval before consequential actions Person
Person: a human reviews or approves before the workflow continues.
Safety and governance
Treated as core architecture, not add-ons
Permissions
- Action limits
- Least-privilege tool access
Credentials
- Secure credential handling
- No secrets in prompts
Oversight
- Approval steps
- Reversible actions
Evidence
- Logging
- Evaluation and monitoring
Hands-on
Hands-on labs
Learners work through realistic scenarios rather than isolated feature demonstrations.
- Create a bounded agent
- Connect an approved tool
- Add retrieval
- Require approval before an action
- Test unsupported requests
- Log and evaluate behaviour
Outcomes
What learners should leave with
- Understand agent architecture
- Design agents with clear permissions and approvals
- Evaluate and monitor agent behaviour
Agents are taught as bounded workflows with clear permissions and measurable outcomes – not as unsupervised digital employees.
Formats and prices
Learn Agentic AI your way
Every format is live and instructor-led, so learners can ask about their own data and situations.
For individuals
Public courses
From $79 per learner
Join a small live online group for one focused course.
- Live online sessions with a practitioner
- Hands-on labs on realistic data
- Course materials to keep
For organisations
Team training
From $600 per day, up to 12 people
A private programme around your tools, data and goals.
- Agenda chosen for your team
- Remote, or on-site from $900 per day
- Optional follow-up clinics
For specific questions
1:1 coaching
$50 per hour
Personal help with your own report, model or query.
- Screen-share sessions
- Work on your real files
- Book single hours as needed
Prices in US dollars, excluding any local taxes. Paid in advance by invoice. Public course groups are limited to 12 learners.
FAQ
Frequently asked questions
Do I need to code?
Basic programming or API experience helps. Labs use Python and a model provider API; setup instructions are sent in advance.
Which platform is used?
The public course uses a vendor-neutral approach. Private cohorts can focus on your preferred platform and APIs.
Need it built? · Datazeb Consulting
Want a controlled AI assistant or agent built for you?
Datazeb builds knowledge assistants and agents with permissions, approvals and monitoring.
datazeb.com